A Unified Masked Jigsaw Puzzle Framework for Vision and Language Models

TL;DR

This paper introduces a Masked Jigsaw Puzzle framework that enhances the robustness of Transformer models against gradient attacks and improves their performance across vision and language tasks by disrupting local spatial information in position embeddings.

Contribution

The paper proposes a novel Masked Jigsaw Puzzle framework that mitigates gradient attack vulnerabilities and boosts Transformer performance in vision and NLP tasks.

Findings

Improves model robustness against gradient attacks

Enhances performance in vision and language tasks

Unified framework for Transformer models

Abstract

In federated learning, Transformer, as a popular architecture, faces critical challenges in defending against gradient attacks and improving model performance in both Computer Vision (CV) and Natural Language Processing (NLP) tasks. It has been revealed that the gradient of Position Embeddings (PEs) in Transformer contains sufficient information, which can be used to reconstruct the input data. To mitigate this issue, we introduce a Masked Jigsaw Puzzle (MJP) framework. MJP starts with random token shuffling to break the token order, and then a learnable \textit{unknown (unk)} position embedding is used to mask out the PEs of the shuffled tokens. In this manner, the local spatial information which is encoded in the position embeddings is disrupted, and the models are forced to learn feature representations that are less reliant on the local spatial information. Notably, with the careful use of MJP, we can not only improve models' robustness against gradient attacks, but also boost their performance in both vision and text application scenarios, such as classification for images (\textit{e.g.,} ImageNet-1K) and sentiment analysis for text (\textit{e.g.,} Yelp and Amazon). Experimental results suggest that MJP is a unified framework for different Transformer-based models in both vision and language tasks. Code is publicly available via https://github.com/ywxsuperstar/transformerattack